WORKFLOW MANAGEMENT WITH AUTOMATIC METADATA
EXTRACTION
As libraries convert materials into digital formats, the need for efficient
workflow management tools will increase. The significant human labor required
for editing, inspecting, correcting, and “tagging” (with appropriate
metadata) digital objects might inhibit libraries and other organizations
from initiating large-scale digitization efforts. By developing the framework
of workflow management tools, semi-automated tools will reduce the resource
requirements for implementing large-scale digitization projects and provide
enhanced search functionality.
This research project will investigate developing an efficient and economical
framework of tools to manage the workflow of large-scale digitization
of musical materials. It aims to support the path from physical object
and digitized material into a digital library repository by providing
effective tools for perusing multimedia elements. The result of the process
will include the audio files; both the images and the text of album covers,
record labels, and liner notes; metadata about the recordings; images
of scores and files in machine-readable format; and a database enabling
the information to be searched and accessed via the web.
An important component of the research is to minimize human intervention
by automatically generating text and metadata from the captured images
using document analysis and recognition techniques. Metadata extraction
constitutes a major source of cost in most digitization projects. In the
case of vinyl records, the possible sources for the metadata are the record
label, album cover, or liner notes. For printed music, the sources for
the metadata may be the cover page and the title page. For music manuscripts,
external sources may need to be consulted. The challenge lies in locating
the appropriate information and extracting it from these different sources.
Since the required data may appear anywhere, automatic extraction of the
data necessitates the deployment of intelligent document analysis.
To implement the specialized document analysis required for this project,
open-source software called Gamera
(Droettboom et al. 2002) will be used. Gamera, developed by the applicant
and others at the Johns Hopkins University over the last several years,
is a toolkit for the creation of domain-specific structured document recognition
applications by domain experts with limited programming experience. It
allows a knowledgeable user to combine image processing and recognition
tools in an easy-to-use, interactive, graphical scripting environment.
The goal of the Gamera system is to leverage the user’s knowledge
of the target documents to create custom applications rather than attempting
to meet the needs of diverse users through a monolithic application. The
applications created by the user are suitable for use in a large-scale
digitization project; they can be run in a batch-processing mode and easily
integrated into a digitization framework. In order to create consistent
names in the metadata, automated name authority control methodology developed
at the Johns Hopkins University (DiLauro et al. 2001) will be added to
Gamera. It is anticipated that this will also aid in the optical character
recognition step, by providing a dictionary of names and their variants.
The infrastructure will provide the sustained support needed for the continual
development and the enhancement of Gamera software.
|
